As someone who doesn't know much about this stuff but has a decent understanding of statistics, I really enjoyed watching this. You pack a ton of very heavy information into a concise, well put together video. Making a video like this (let alone know all of this) at 13 years old is very impressive! Keep up the good work!
Thank you very much for your great work! A quick note: If I remember correctly, the R2 *score* can be arbitrary negative, whereas the plain R2 value is in a range between 0 and 1.
Very very clear explanations, you seem to explain the things almost every other RUclipsr misses, thanx a lot this has helped me more than any other video in explaining the steps to implementing a well defined model.. thanx and keep those videos coming
Linear regression model is always a input layer straight to the output layer ? i seen once people use 2 dense layer to do a regression model. i just wondering can i add more layer to do a linear regression ?
i execute you code step by step but in this command line to calculate the accuracy i become error ValueError: y_true and y_pred have different number of output (1!=100) # Calculates and prints r2 score of training and testing data print("The R2 score on the Train set is:\t{:0.3f}".format(r2_score(y_train, y_train_pred))) print("The R2 score on the Test set is:\t{:0.3f}".format(r2_score(y_test, y_test_pred)))
Thank you so much for the video. I have learnt a lot from young man like you. But I encounter a problem, each time after retraining the model, the result is different even though parameters stayed remain, so it' quite problematic when I want to test the influence of parameters in the model. Is there any way to improve this?
What do you mean in particular by adding 13 outputs to the first layer at 15:30? As far as I understood, the 13 mean that you have 13 neurons in that layer. What is not clear to me is the fact that on the one hand you have 3 input features but also 13 nerons that are in the first layer. How does that work. I thought you have to have one neuron for every input feature?
Cristal Clear, excellent videos!!! I could almost do everything. But Jupyter does not show OUTs... When I run all cells or one by one, out[] is not showed... According what I was searching in forums, faqs etc, many other people have the same problem...
Thanks! I have experienced the same issue before. While it may not be the most elegant solution, I often just create a new Python file and paste all my code into it. Let me know if this works!
It seems that the activation function needs to be 'tanh' instead of 'relu' to have negative values as output. If anyone has some more information on how to implement tanh in the code, please share :-)
Big props to you man and your dad! You are 13y old!!?? And i am here still trying to figure how to learn python and programming for start
At 13, the smartest programming I could do was "int a=1, int b=2, print(a+b)". You rock sir!!
Sincere thanks from a 42 years old dog trying to learn new tricks. What my lecturer couldn't explain, you did so in 18 mins
13 Years old, I am impress. thank's that was awesome!
As someone who doesn't know much about this stuff but has a decent understanding of statistics, I really enjoyed watching this. You pack a ton of very heavy information into a concise, well put together video. Making a video like this (let alone know all of this) at 13 years old is very impressive! Keep up the good work!
Thanks! I appreciate it.
Kudos to you young man. You should be very proud of yourself and keep up the good work!
Thanks!
great explanation bro, easy to understand you should make videos more frequently to help keen learners like us
Good job buddy! as a ml engineer I can say that this video is great and very informative.
Thank you!
Thank you very much for your great work! A quick note: If I remember correctly, the R2 *score* can be arbitrary negative, whereas the plain R2 value is in a range between 0 and 1.
This is so underrated. Amazing video and very educational
Very very clear explanations, you seem to explain the things almost every other RUclipsr misses, thanx a lot this has helped me more than any other video in explaining the steps to implementing a well defined model.. thanx and keep those videos coming
Thanks!
Holy.... This is a well made video! If you are still active in Python please do more content like this. This is gold...
Good video covers a lot of the parameters, ihavent find this detail anywhere else in youtube.
This is a very well made and informative video. Kudos to you :)
Thank you
Great video. Interesting and educative
Amzing video dude... Nice explanation. Pls continue more videos like this. Thankyou.
Wow, amazing video ! Everything that we need to understand the differents concepts is really well explained. Congrats !
High quality video with lots of information, thank you 🙏
Thanks!
Wow, your videos helped me a lot to setup my case. Your explanations are very clear, keep up the good work! :)
Excellent work....very helpful 👌
You are so brilliant man can't wait for new videos😍
Thanks!
Linear regression model is always a input layer straight to the output layer ?
i seen once people use 2 dense layer to do a regression model.
i just wondering can i add more layer to do a linear regression ?
This is absolutely fantastic, thanks!
You are very welcome!
plz make more videos on keras as soon as possible keeep it up
Dude you're great - keep the videos coming!
i execute you code step by step but in this command line to calculate the accuracy i become error
ValueError: y_true and y_pred have different number of output (1!=100)
# Calculates and prints r2 score of training and testing data
print("The R2 score on the Train set is:\t{:0.3f}".format(r2_score(y_train, y_train_pred)))
print("The R2 score on the Test set is:\t{:0.3f}".format(r2_score(y_test, y_test_pred)))
Amazing video! Keep up the great work!
Great video. Please post more vids :)
WOW YOU JUST GREAT DUDE , GREAT EXPLANATION
Thanks!
Thank you so much for the video. I have learnt a lot from young man like you. But I encounter a problem, each time after retraining the model, the result is different even though parameters stayed remain, so it' quite problematic when I want to test the influence of parameters in the model. Is there any way to improve this?
Great video. Keep uploading videos
This is good tutorial
Thank you!
Thanks! bro, how do you visualize de NN structure?
What do you mean in particular by adding 13 outputs to the first layer at 15:30?
As far as I understood, the 13 mean that you have 13 neurons in that layer. What is not clear to me is the fact that on the one hand you have 3 input features but also 13 nerons that are in the first layer. How does that work. I thought you have to have one neuron for every input feature?
The input layer has 3 neurons, and the second layer has 13 neurons. That’s what I meant when I said the first layer had 13 “outputs.”
@@hwhd ok but what does the 13 mean in the very first layer?
Hi, very well explained. How did you increase the degree of the regression? Is it just by passing more hidden layers?
Good Job! Keep up the great work!
Thanks!
Did you use Selenium to crawl through youtube and scrape the data? I've been wanting to do the same thing and Selenium seems like a good way to do it.
Actually, Beautiful Soup was used to collect the data
@@hwhd Oh okay. So did you do manual queries on youtube, then scrape those pages with BS?
Yes
I guess you need to scale before splitting train and test data, in case they have different distribution
This is fantastic!!
Thanks!
Well done bro, you rock dude
Great video thank you.
Thanks!
great tutorial, thanks
You are very welcome.
Cristal Clear, excellent videos!!!
I could almost do everything. But Jupyter does not show OUTs... When I run all cells or one by one, out[] is not showed...
According what I was searching in forums, faqs etc, many other people have the same problem...
Thanks! I have experienced the same issue before. While it may not be the most elegant solution, I often just create a new Python file and paste all my code into it. Let me know if this works!
Great Video! BTW you didnt define your weights and biases??
Amazing video !! Could you also make videos on how to use a web crawler to make your own dataset
Thanks :)
You can also use RUclips's API for data extraction
Is web scraping and web crawling the same?
I am also interested to know about the youtube web scraping.
Does tf have a function to visualise the model?
Really informative! Could you please share the code for your web crawler if possible ? Keep up with the good work .
You think i have any idea what the fudge is going on
Smooshed ikr look at this loser
Can you reverse it? Give it the number of views you want and have it give you a topic or keywords to game the searching and home screen algorithm?
It is nearly impossible to “game” the RUclips algorithm, so no.
@@hwhd even getting a list of tags that could bump you up in the search would be worth it
Make vedios about vgg and resnet
what kind of neural network is this?, RNN, CNN?
Neither. It is a plain old feedforward neural network.
Querreies
Does this model work with X and Y values that are both positive and negative? Does any changes need to be made in order to do that?
It seems that the activation function needs to be 'tanh' instead of 'relu' to have negative values as output. If anyone has some more information on how to implement tanh in the code, please share :-)